Hemse Stave Revisited BY GUNNAR ALMEVIK & JONATHAN WESTIN

Abstract During a restoration of the Romanesque church in Hemse in 1896, the remains of a stave church were found as reused floor tiles. The discovery was important at the time, providing new information to a prestigious research field with few sources of knowledge. Today, the remains of Hemse are esoteric and inaccessible for scholarship. The stave church material in the present museum context seems to have one function, to communicate an age value. The leading question in this article is, what more could we retrieve from this old archaeological material? We may agree that the museum’s archaeological collections and the stave church remains are valuable sources, but for what new kind of knowledge? This article presents the process and outcome of an in-depth examina- tion of the material remains and archival records of Hemse stave church. The aim is to develop or revise how this wooden church may have been constructed and appeared both outside and inside when it functioned as a building. The research method uses three perspectives that give access to different paths of knowledge: a discursive perspective, a forensic perspec- tive, and a dwelling perspective. The research results are contextualized in an interactive model of Hemse that provides a visual experience that gives a sense of the stave church as a real place and not just a theoretical space. The results are grounded on empirical evidence but also on the intellectual discourse of which it is a product. The reconstruction is less of a static representation of our knowledge than a simulation or a research laboratory through which hypotheses can be tried and both researchers and the public can be engaged in a dialogue.

Introduction

During a restoration of the Romanesque Antiquities expert Emil Ekhoff, who arrived at church in Hemse in 1896, the remains of a the construction site the very next morning. stave church were found as reused floor tiles. Ekhoff collected in all 67 pieces of the wooden The local contractor Nils Pettersson informed church that were bought from the parish the Royal Academy of History, Letters and and brought from the island of to

LUND ARCHAEOLOGICAL REVIEW 23 | 2017 7 Figs. 1–2. Reconstruction drawing and context of material evidence by Olof Sörling, published in Emil Ekhoff’s research opusSvenska stavkyrkor 1914–16. The construction is of oak with vertical wall timber between a sill and a top plate. All “staves” have been cut in the lower part, but a groove in the sill indicates that the staves were jacked in the sill but continuing into the ground. The wall staves have a convex outer form with decorative concentric rings in a cross shape at the top and a characteristic carving at the plate. Ekhoff’s interpretation borrows forms from the Norwegian stave churches of Urnes and Hopperstad, and faraway Greenstead church as well as nearby Garde. ATA.

the National History Museum in Stockholm When Emil Ekhoff published his (Ekhoff 1914–16, 23). At the time there were, opus Svenska stavkyrkor (“Swedish Stave and still is, only one preserved stave church Churches”, 1914–16), Hemse was brought to in , namely Hedared in Västergötland. the fore as the main example. The National Hemse was not as splendid as the Norwegian History Museum’s “artist” Olof Sörling national treasures Urnes or Kaupanger, but provided detailed documentation drawings, still with archaic forms and traces of refined and a reconstruction that has been frequently ornamentation of which Hedared had none. reproduced in succeeding research to illustrate The discovery of Hemse was important a Swedish stave church. The reconstruction at the time, providing new information to a drawing of Hemse appears in the mandatory prestigious research field with few sources of folder with the church history in Hemse, knowledge. The investigation caused an intense in the monograph about Hemse church in debate with the Norwegian scholars on how the national inventory Swedish Churches to interpret and denominate the construction [Sveriges kyrkor] (Lagerlöf & Stolt 1969), (Forntidsminneforeningens årbok 1898, 48ff.). In in various exposés of Gotlandic churches 1910 the National History Museum produced (e.g. Lagerlöf & Svahnström 1991), in a new exhibition on Romanesque church Anders Åman’s outline Träkyrkor i Sverige art, where Hemse stave church was staged (“Wooden Churches in Sweden”, Åman & in centre of display. Decorative parts from Järnfeldt-Carlsson 1999) and Claus Ahrens’ the west portal were fitted in a reconstructed overview of Die Frühen holtzkirchen Europas stave church facade and used as entrance to a (“The Early Wooden Churches of ”, diorama of a church space (Bengtsson Melin Ahrens 2001). Though Ekhoff’s investigation 2014, 64). Bernhard Salin, later to be director and illustrated reconstruction of Hemse has of the National Heritage Board, personally been referred to, reproduced and also in arranged the exhibition. some respects criticized, the sources has not

8 GUNNAR ALMEVIK & JONATHAN WESTIN been thoroughly revisited and revised. The in a decorated top plate with extended corner reconstruction drawing has become what Jean joints. There are two small windows placed Baudrillard calls a simulacrum, a conjured high in the south facade and three more copy that supersedes the original as reference openings in a clerestory-like elevation in the to historical truth (Baudrillard 1994). sectioned slope of the roof. The doorway in Today, the remains of Hemse are esoteric the west portal is open, but offers no glimpse and inaccessible for scholarship, distributed in of the interior space. On what evidence is this various spaces for deposit and display. In the reconstruction based? What would we see if National History Museum, the reconstructed we could step over the threshold of this tall, western portal is still in the exhibition, but narrow doorway? now placed against a wall and lifted slightly Our research interest in the appearance from the ground hanging like a painting. of a single monument may seem simple, but The entrance is used as a frame for a wooden we will argue that an evidence-based visual sculpture of a sitting bishop. The display is a reconstruction of a particular historic artefact contradiction in terms of a building, providing may be a scholarly challenge. Our starting no path to experience the exterior or interior point is the material remains of Hemse, but church space as something real and tangible the interpretations of this incomplete source for the human hand. The archaeological require a bricolage, where a diverse set of sources remains in this museum context seem to are consulted and triangulated (Almevik have one function, to communicate an age 2012; Mogren et al. 2009). Furthermore, value. The leading question in this article is, an object like Hemse stave church has to be what more could we retrieve from this old examined in its twofold present form, as both archaeological material? We may agree that an historical object and a research object. We the museum’s archaeological collections and want to ground our research on empirical the stave church remains are valuable sources, evidence but also on the intellectual discourse but for what new kind of knowledge? of which it is a product (see Mol 2002, ix). Hence, our solution is a method design through three perspectives that give access to different paths of knowledge: a discursive Research aim and method perspective, a forensic perspective, and a The aim of this research is to reinvestigate dwelling perspective. The method design also the archaeological building remains from recurs in the arrangement of the article. Hemse and develop or revise how this The first two sections present a discursive wooden church may have been constructed perspective aiming at eliciting relations between and appeared both outside and inside when it the structure of narratives that explain the functioned as a building. The reconstruction research object and the motives and methods drawing of Hemse (Fig. 1), signed by Sörling in the production of history. It is a kind of and published by Ekhoff 1914–16, provides archaeological excavation endeavour with a perspective of the church exterior from reference to Michel Foucault, to unfold the the south-west. The reconstruction drawing layers of interpretations that have intersected bestows distinct and seemingly inviolable Hemse during more than a century as both a qualities on history. The construction of the research object and a museum object (Foucault nave and chancel has round corner staves with 2002). The first section focuses on the process extensive decoration. The staves or wall planks of discovery and formation of a discourse of are vertical, anchored in the ground and fitted Hemse as a conspicuous example of a Swedish

LUND ARCHAEOLOGICAL REVIEW 23 | 2017 9 stave church. The second section focuses on connects cultural historical interpretations to succeeding excavations of two stave churches the building’s effect on our senses. Through in the nearby locations Silte and Garde, and the VR reconstruction, the space of Hemse can the scientific construction of a stave church be experienced as a place, a “humanized” space typology where Hemse represents a transition that is tangible for the human hand (Norberg- from earth-embedded staves to a ground plate Schulz 1980). The affordances from the stave construction. agential meshwork that Hemse stave church The second perspective of the article looks constitutes is examined through various at the wooden construction as a primary source aspects that make the space a place, such as of knowledge. This forensic perspective takes us artefacts, order, character, light, and time on a journey to trace both the whereabouts of (Ingold 2007; Gibson 2015 [1979]; Dmitruk the archaeological remains, and the qualities 2017). Hence, following Maurizio Forte, in form and substance that may lead to new who proposes a move from basic interaction interpretations. The forensic perspective refers to feedback, behaviours, and embodiment to a criminological investigation method (2014), we approach the act of reconstruction aimed at bringing evidence to a specific act with a desideratum to explore a possible turn of significance (Almevik 2012; Almevik & from an inviolable visualization process to an Melin 2016; Almevik 2017). The approach open-ended simulation process. follows what Carlo Ginzburg calls a “paradigm of clues”, using a semiotic pragmatism to describe the seemingly insignificant detail “On the correct use of the term and re-enact a system of meanings to uncover the traces of events which the observer stave church” cannot directly experience. It is a general Emil Ekhoff presented the first results archaeological premise: “When causes cannot from the investigations of Hemse in Ny be repeated, there is no alternative but to Illustrerad Tidning already in 1896. He clearly infer them from their effects” (Ginzburg & established the differences in the construction Davin 1980, 23). The forensic reinvestigation of Hemse and the Norwegian stave churches: of the stave church material is not unlinked “The properties of the plates in the frame to previous research, but rather a dialogue to clearly show that corner posts, which are test, attest or contest previous observations an important part of the Norwegian stave and statements. churches, have not existed in the Hemse The third perspective leads to the result that church” (1896, 382). To advocate the use of is a simulation of a multi-sensory experience of the term stave church despite the difference Hemse stave church. The dwelling perspective from the Norwegian wooden churches, he refers to Tim Ingold’s investigation of skills derives the term stave etymologically as a and argumentation for an eco-dynamic traditional term for raised wood, and suggests approach, “which situates the practitioner, a definition of the type as: “a building whose right from the start, in the context of an active walls consist of vertically raised timber, not as engagement with the constituents of his or the corner-jointed building with horizontal her surroundings” (2011, 5). The means in wood” (1896, 374). our case is not a full-scale physical processual Ekhoff’s use of the term stave church caused reconstruction, but a Virtual Reality (VR) a reaction among contemporary Norwegian reconstruction that both encompasses a step- scholars, and the first to defend “the correct by-step enactment of a building process, and use of the term stave church” was Herman

10 GUNNAR ALMEVIK & JONATHAN WESTIN Schirmer, at the time chair of the Society concludes that “Hemse raised timber church for the Preservation of Ancient Norwegian would probably be an adequate denomination” Monuments [Fortidsminneforeningen] and (Fortidsminneforeningen 1899, 49). later director of the National Heritage Board. The debate has a gravity with more at The debate was published in the Society’s stake than academic adequacy and prestige: it yearbook in 1899 since Ny Illustrerad Tidning concerns the branding of stave church as an dismissed the Norwegian objections. “Stave immaterial property with strong implications church”, Schirmer declares, “is a modern for national identity. Leif Ancker has shown word, clearly not known in our medieval in recent historiographical research how the literature” (Fortidsminneforeningen 1899, Norwegian stave churches were “discovered” 48). He rejected Ekhoff’s definition of the and reevaluated as national monuments in the term stave church as a raised timber wall, and 19th century, and how this patrimony at the proposed that it instead refers to the particular time of the debate was institutionalized and constructional features in the Norwegian internationally recognized in the European tradition: “The only connection concerning academic community (2016). No Swedish churches where staves exist, is corner staves, curator should infringe on this prosperous namely the very constructional part of the heritage. Nevertheless, Echoff stuck to the building that you have shown has not been concept and used the brand to entitle his opus present in the construction of Hemse church, “Swedish Stave Churches”, where Hemse as well as the stave plates [staflægjur], the church is put forward as the grand example. horizontal wood that rests on top of them, One significant detail and trace of this are supported by them and connects them” debate is the construction of the corners of (Fortidsminneforeningen 1899, 48). Schirmer Hemse church. In the reconstruction drawing

Figs. 3–4. To the left, the first interpretation of the construction with vertical wall timber without a corner post presented in 1896 (Ekhoff 1896, 376). To the right, the final reconstruction from 1914–16 with a construction with a structural corner stave with ornamentation derived from Urnes. ATA.

LUND ARCHAEOLOGICAL REVIEW 23 | 2017 11 there are round decoratively carved corner staves, very similar to the oldest preserved parts of . How did Ekhoff’s interpretation of the archaeological material change, from a statement that it “clearly shows that corner posts […] have not existed” to this appearance very similar to idea of what the oldest version of Urnes might have looked like? The time lapse between the discovery of Hemse and the publication of the full investigation encompasses almost two decades. Ekhoff cooperated with Sörling, not only with the final illustration drawing, but also during the process of interpretation. The standard method was that Sörling made a kind of contour sketches in planear-projected elevations of the stave church parts at a scale of 1:25. Ekhoff made notes on these contour sketches and used them as templates in reconstruction experimentations (Fig. 5–6). The preserved archival documents from the investigation of Hemse clearly show that the main concern was the Norwegian connection with the corner stave (Fig. 7). The accumulated data from the decades of investigation is extensive, and the final product discloses the problems Ekhoff had in managing the research material. Though the archive contains page after page with minute measurements in millimetres of each and every stave, notes on every gutter and notch, the first published examination of the material from the year of discovery is more accurate than the extensive descriptions in the concluding research almost two decades later. In the final presentation of Hemse the measurements on the ground plan of the church are not presented except through a ruler accompanying a drawing of the sill showing a length of the nave that is half a metre longer than it actually is: 8.4 metres

Figs. 5–7. Drawing from Ekhoff’s personal archive, indicating the work process. ATA.

12 GUNNAR ALMEVIK & JONATHAN WESTIN instead of 7.87 metres. Furthermore, though only one full-length chancel sill was listed in the record of collected items from Hemse, in the final presentation one more sill appears, now providing full information about the church plan. It seems that Ekhoff got detached from the primary sources, and made his final publication partly based on misinterpretations of his own fragmented representations.

Fig. 8. Drawing by Kristian Bjerknes of the An informative imprint of a development of types of stave constructions, stave church from earth-embedded to sill-carried walls. In this scheme Hemse represents a transitional type. Ekhoff remained an authority in the field of Drawing: Kristian Bjerknes (1972, 99). Swedish stave churches, and the publication of Hemse had an impact on succeeding research. The debate on the stave church as, according to her, the construction is not brand died down, and in the encyclopaedia original (Boëthius 1931, 49). The inner of the Nordic Middle Age (Kulturhistoriskt corner in Greenstead has been cut out in a lexikon för nordisk medeltid), the authors of the recent building phase. Furthermore, Boëthius headword “stave church” briefly conclude that discusses the stability of a construction like there exist two parallel definitions of the word: Hemse with raised timber between a sill and in Swedish referring to standing wall timber, a top plate, without firmly connected corner in Norwegian meaning the corner stave in a posts or braces, of which there are no traces. roof-bearing framework (Bjeknes 1972, 95; Her conclusion is that there ought to have Anker 2016, 17). Hemse is used not least by been inner earth-embedded posts connected Norwegian researchers like Kristian Bjerknes to the roof construction through the truss (1972, 99), Roar Haugli (Haugli 1976, 192) (Boëthius 1931, 121). and Peter Anker (Anker 1997) as an example The questions about both the stability of a transitional type, between the so-called and the missing parts of the wall staves palisade construction with earth-embedded were elucidated in the early 1970s when walls of half timber, mainly represented by St an informative imprint of a stave church Maria Minor in Lund, and the stave church was discovered in Silte. The imprint in “proper” like most of the preserved Norwegian this neighbouring parish seems to be like stave churches with -bearing corner staves Cinderella’s shoe to Hemse. In 1971 Harald placed on a sill (see Fig. 8). Olsson was commissioned to conserve the Scholars contemporary with Ekhoff, mural paintings in the Romanesque church in however, raised critical questions, particularly Silte. During work on the inner wall in the about the reconstruction. The art historian nave connecting the chancel, he discovered Johnny Roosval suggested that Hemse had a stains from tar with the contour of a possible hipped roof like historic churches in the Baltic stave church roof. The hypothesis was that a region, and thus no gables (Roosval 1924). wooden stave church had been standing while Gerda Boëthius questions the corner post a new stone church was being built. and the reference to the church in Greenstead An excavation of the church foundation

LUND ARCHAEOLOGICAL REVIEW 23 | 2017 13 Figs. 9–10. To the left, excavation of the foundation of an earlier stave church in Silte parish. Cut- outs of the vertical timber staves have been positioned to visualize the placement of the staves. Photo: Waldemar Falck, January 1972. ATA. To the right, documentation of the contour line in the nave of the stave church, showing an inclination of approximately 20 cm from the ground to the top plate. SfM documentation by Gunnar Almevik.

was initiated, led by the archaeologist Gustaf ground. The boulders fixed the wall staves in Trotzig. The stone church nave was fully place. The contour of the roof in Silte provides excavated and a foundation of a smaller precise information about the roof height church was uncovered. There were no and the inclination angle of an absent stave significant findings of wooden construction church. In the stain from the roof tar we find material, except for a piece of a wooden roof no hint of a raised mid section or sectioned shingle, but the properties of the foundation roof fall. It is a plain tall raised gable roof. were peculiar. Just outside a stone base for There are two properties of the find in the sill there was a gutter about 45 cm deep, Silte that are documented but not further and further outside larger boulders, probably discussed. Firstly, the walls have a notable to keep the wall in place. The floor was made inclination inwards. Our own measurements of planed lime mortar on top of a deep layer show that a cross-section of the nave at the of limestone chippings. The mortar floor was level of the sill was about 40 cm wider than placed on a level with the stone base and the same cross-section up by the top plate. continued at a level between the nave and Claus Ahrens depicts the leaning walls in a chancel. reconstruction sketch (Ahrens 2001, 219), Through the Silte excavation, the absent but what does this do to the appearance of the part of the staves and the question of stability church? It is obviously intentional and not a in Hemse found a plausible explanation. The matter of decay. staves probably clung to the slanting cut in The second perhaps overlooked detail is the sill and continued down the gutter in the the traces of a central column. According to

14 GUNNAR ALMEVIK & JONATHAN WESTIN Fig. 12. Garde excavation plan from 1968 (Trotzig 1970, 7) showing the earth-embedded staves, a foundation from a baptismal font and (hole no. VII) a hole from a possible earth- embedded centre column, 30 cm in diameter and 50 cm deep (ibid., 8). Fig. 11. Gustaf Trotzig’s interpretation of the wall staves, jacking the sill and continuing into a gutter (Trotzig 1972, 84).

Gustaf Trotzig’s description of the excavation construction, is not possible to attest by the in Silte, there is an earth-embedded post- archaeological record. The excavation sheds hole in the former church nave, in centre to light on some details of the interior plan: a the west, 15–20 cm in diameter and 50 cm large foundation of a baptismal font with a deep (Trotzig 1972, 75). The post-hole is not drainage arrangement is placed in centre with depicted on the documentation drawings. an offset to the west. In a possible epicentre of This may have been a structural element of the the church, behind the basin, there is also a wooden church, as Gerda Boëthius suggests. hole from an earth-embedded central wooden We know several Gotlandic churches with a column. central stone column structurally related to the vaults. Wooden central columns also exist in a number of Norwegian stave churches Examining 67 pieces of historic with connection to the roof construction. We discern that this central column is an object wood of significance. Significant parts of Hemse stave church, three In the nearby church of Garde yet decorated pieces from the western portal and another stave church was discovered in the seven staves, are used in the reconstructed Romanesque stone church nave in 1968. facade at the National History Museum. The excavation, also led by Gustaf Trotzig, Another six staves are used in a wooden revealed a church plan between thick earth- lapidarium in the Museum of Gotland. The embedded posts (Trotzig 1970). If these only preserved stave with a connection joint posts carried a ground plate for the wall to the chancel is deposited, unmarked, in structure, or if the wall staves had a palisade the museum storage outside . A carved construction, separated with the roof carrying detail, supposedly a gable end decoration,

LUND ARCHAEOLOGICAL REVIEW 23 | 2017 15 is located in the cellar of the main building of the History Museum. The rest of the 40 pieces are located in the Swedish Museum Service’s grand store in Tumba. Ten pieces are missing of the presumed 67 items in the History Museum’s inventory: eight fragments of staves, a piece of fir wood, and a chancel sill that might never have existed (SHM 10232– 5). The two benches that were found during the restoration of Hemse in 1896, hidden as support for other benches, are stored separately from the rest of the stave church material. All construction parts have been reinvestigated, documented and processed into digital 3D representations through Structure from Motion (SfM). Using photogrammetry, SfM computes a three-dimensional structure from a large amount of photos captured systematically (see de Rue et al. 2013). The investigation involved a general survey of the surfaces with a tactical flashlight to elicit traces in the wood. Comprehensive measurements were taken but no manual drawings were made. Erik Hansen argues that the manual measurement and documentation drawing is not just a means of representation but a mode of investigation, providing “a dialogue” with the monument (Hansen 2000). On the other hand, the gain with the digital documentation is the opportunity to continually reinvestigate, check, and control data. A manual drawing Figs. 13–15. Traces of tool marks indicating a of an artefact, such as a nine-metre long top social division of surfaces. plate, may generate about 150 measurements, and a digital documentation through SfM may documentation contests core measurements generate a point cloud containing 32 million of the stave church published by Ekhoff measurement points. As representations rich 1914–16. The inaccurate measurement of the in detail, the digital models eliminate a level of length of the sill has been taken for granted abstraction that drawings introduce. In Ekhoff by subsequent researchers. When for instance and Sörlin’s work process, the translation from Erland Lagerlöf and Bengt Stolt compared physical artefacts into inscriptions meant a the church space in Hemse with Eke and Silte loss of contact with the originals where errors they reproduced Ekhoff’s errors (Lagerlöf & were easily introduced. Stolt 1969, 185; Lagerlöf & Stolt 1974, 465). The reinvestigation of the remains from Secondly, we observe a conscious grading of Hemse gave new information. Firstly, our the wooden surfaces. The hidden surfaces

16 GUNNAR ALMEVIK & JONATHAN WESTIN on the sill are roughly hewn in the peculiar early medieval technique sprätthuggning (Fig. 13), while the surfaces close to the human hand are smoothly planed (Fig. 15). Visible construction parts in distant places like the top plate and the roof (there are two parts of possible members of the truss) have a semi- smooth surface carefully hewn with a broad axe (Fig. 14). Our investigation affirms the critique of Ekhoff’s reconstruction using rounded corner staves, which made Hemse resemble Urnes I. Observations that point in another direction might have been overlooked. There is one stave in particular, with a pattern subtly distinguished from all other staves through the inclusion of two instead of one formation of a cross-shape made out of concentric rings, that could be the one located in a corner joint. Unfortunately the sides have been hewn and planed, leaving no trace of the possible corner joint. Furthermore, there are no traces of any windows in the preserved staves. This is however evidence ex silentio. Erland Lagerlöf and Bengt Stolt argue that any penetrated staves were unfit as floor tiles and sorted out when reused (Lagerlöf & Stolt 1969). Were there no Figs. 16–17. To the left, the groove in the sill windows at all? We cannot tell for sure, but to connect the wall staves continues to the end. we question the habitual interpretation that There are no traces of a corner stave. The small a building is defined by the property of light hole may be a secondary nail to hold the pieces through windows. The early Christian religion together after re-erection. To the right, traces of presumed original method to fasten the wall revolved around the light ritually provided staves in the top plate with exterior wedges, mainly through oil lamps or candles. and secondary skew dowels mounted from the The reinvestigation also confirms many interior. of the observations that Ekhoff made. The preserved staves from the inner wall between nave and chancel imply a narrow opening, possibly a doorway as suggested by Ekhoff. The west portal entrance door opened inwards Imprints in the tar and nail holes along the and was mounted on the left side of the inner sides of the staves are a lead to a cover strip wall. There are several leads that inform that between the wall staves, possibly to cover the building was either repaired, or more emerging gaps from decay or subsidence. probably moved and re-erected. There are When Ekhoff publishedSvenska traces of both primary wedges and secondary stavkyrkor, remains of wooden churches were dowels to fix the wall staves in the top plate. known in , and Lye

LUND ARCHAEOLOGICAL REVIEW 23 | 2017 17 (Ekhoff 1914–16). Later discoveries have what is notable is their decorative features: been made in Eke, , Dalhem, Sproge, the standard shingle is pointed, often with Garde and Silte. The primary interpretation an engraved relief and with a visible iron potential occurs in dialogue with the later nail (Utas 2008; Lagerlöf 1978). Another stave church discoveries in Silte and Garde. In decorative feature is the concentric rings, the light of the findings at these sites we can say same as found in Hemse. that the staves in Hemse were jacked in the sill but continuing down in the earth, giving stability to the construction. Large boulders were placed against the outer walls as support An interactive diorama but also with a clear visual effect. The walls The third perspective in this article seeks leaned inwardly, which we assume is partly a dwelling perspective where the results of due to a thickening of the lower part of the the research are compiled and graphically staves. The 4.35 metre inclined palisade-like presented as an alternative reconstruction to walls and the steep gable roof, well over seven the one Sörling and Ekhoff presented in 1914– metres to the ridge, gave the small church a 16. Using the SfM documentation of the 57 tower-like silhouette. The door, 3.3 metres pieces as templates for a digital reconstruction high and just 80 centimetres wide, intensified model (Fig. 18), each individual part of the lofty impression, and lent the construction the structure has been modelled separately, a kind of grandness in spite of its diminutive piece by piece, like a simulation of a real dimensions in our retrospective lens. The roof construction process. While the main was most likely covered with wooden shingles. measurements are taken from the primary Evidence of shingles from the 12th century material, principles have been derived from are found in several Gotlandic churches, and analogies to other archaeological material

Fig. 18. The digital reconstruction of Hemse Stave Church in Autodesk Maya.

18 GUNNAR ALMEVIK & JONATHAN WESTIN and standing buildings. In this sense we are information, incorporating as it does both conjuring up a church in much the same hypothetical and factual material. Neither is way Ekhoff and Sörling did. However, our such a representation uncommon, as most, analogies are brought closer to the location of if not all, physical dioramas that incorporate Hemse. authentic remains are bound to be made up The aim of the reconstruction is to provide of materiality from different time periods. a visual experience that gives a sense of the stave However, to arrive at a more accurate church as a real place and not just a theoretical simulation of lighting conditions and the space. In addition to accurate construction appearance of the interior space at a time principles and architectural forms, this when the church was in active use, as a place of also requires paying meticulous attention worship, we edited the textures obtained from to textures and light. A critical question in the archaeological finds. This was intended the process of texturing is what stage of the to give them – and in the process also the building’s life span is to be simulated. The simulated stave church – a less weathered textures are obtained from the archaeological appearance. finds and can as such be considered authentic. To build our simulation of this space, and However, applying them unedited to a digital make it into a place, we assembled the model construction that reconstructs the church in Unity, a visualization engine by Unity as it might have appeared when newly built Technologies that also functions as an archive results in a hybrid representation that brings for heterogeneous digital material. Sound together both processed and unprocessed recordings, 3D models, photographs, and data; the digitally reconstructed translation of drawings can here be arranged in an interactive the wooden remains, and the surfaces of the scene. Organized as a place, the interactive wooden remains as they appear today. virtual reality scene resembles in principle a Such a hybrid is not wrong per se, as traditional museum diorama. The diorama is it constitutes a knowledge model rich in a setting in which artefacts or specimens are

Fig. 19. The textured model assembled in Unity.

LUND ARCHAEOLOGICAL REVIEW 23 | 2017 19 mixed with artificial constructions and matte in its composition and appearance, albeit painting to create a scene that communicates more hypothetical. In an effort to put the to a context within the confinements of a focus on the structure of the stave church, display. Hemse stave church has itself been in our interactive scene we have opted for part of a museum diorama, then as a backdrop an empty and featureless landscape, with to an arrangement of early medieval artefacts low non-descript vegetation. The light in the and produced to contextualize their relations scene simulates a clear day at the beginning in a church space. In our use, a virtual diorama of March, around three in the afternoon. is a technique by which to contextualize an These light conditions are chosen to offer a artefact or milieu by bringing it together with well-lit exposition of both the west and south digital representations or reconstructions of facades. However, we are aware that the materiality that have been spread out in both composition of this scene is in no way neutral, time and space (Westin & Almevik 2017). and may lead to faulty assumptions about the Working with photorealistic and immersive structure’s relation to a built environment environments requires an understanding that and the surrounding landscape, as well as such visualizations are easily read as complete communicating a problematic idealized image (Westin 2014). The empty spaces created of the structure’s appearance. from missing materiality, left out of the scene While the inside of the stave church due to uncertainty or lack of information, constitutes a confined space, when one become equally defining parts of the scene approaches it as a liveable place it opens up as the materiality included. Hence, there is to an even larger number of interpretations a conflict between the desire to create a well and uncertainties. In the confined space of researched space only referencing what we the church interiors we are able to simulate know, and the desire to create a place realistic several aspects that make it into a place. In

Fig. 20. The highly polished figurine in the shrine, here represented by the Appuna madonna, glistens as it reflects the light of the oil lamps; the reliefs of the font deepen and the water reflects the light upon walls and ceiling. Shadows, all moving from the flickering lights, become part of the dwelling: the carried light makes the roof truss and rafters come alive as we move through the nave, and the shadows cast by the font and the central column stretch out before us, tracing the outlines of the room.

20 GUNNAR ALMEVIK & JONATHAN WESTIN Fig. 21. The simulation offers opportunities to deconstruct the reconstruction and hold it accountable to the source material. our simulation we start with only two light or lit. Finally, there is the possibility to light a sources: the piercing daylight coming through fourth oil lamp, this one carried by the person the high door opening, and the much more moving through the rooms. By being able to discreet light of the small window in the manipulate light sources, their effects on the chancel. In addition to these, three stationary architectural space become apparent, but also artificial light sources simulating oil lamps their importance for the reception and uses of are available. These invite a focus on three the artefacts. important nodes in the nave and chancel – the As the light and shadow play are such font, the shrine, and the chancel altar table – prominent aspects of the interior space as a and they can independently be extinguished place, interpretations of the architectural

LUND ARCHAEOLOGICAL REVIEW 23 | 2017 21 structure and the furnishing gain an even now began to be built of stone instead of greater importance as these have a profound wood. effect on the lighting conditions: lime-washed walls reflect the light differently from uncoated Cultural-historical interpretation cannot fir wood, and different interpretations of convey the meanings of this building, the roof framework of trusses and rafters beyond this rather poor narrative. We would pick up the light in different ways. may assume that the early Gotlandic stave Furthermore, the addition of windows in churches were commissioned by magnates the south facade, as Ekhoff proposes, would and used within a family or close community dilute some of the controlled – for the liturgy (Andrén 2009). Gunnar Smedberg interprets possibly essential – effects the oil lamps offer. through provincial laws that a process of Hence, the simulation of the interior communalization of the churches was offers a space to try out different aspects of initiated during the late 12th century with what makes this space into a place, and lets the formation of parishes and the church’s us experience it as a real room. It is as a place, development and defence of a tax system where artefacts, order, character, light, and (Smedberg 1973). Hemse, which is dated time all come together, that the effects of by dendrochronological analysis to 1145 the central column as a room divider can be (Bartholin 1998a,b), may have been a experienced and analysed. The font and other magnate’s church, which during its life artefacts can here be assessed in relation to course was used or reused as a parish church. a room, their scale finally regaining context. However, these scholarly narratives are Moreover, being a complete model of the stave concealed beyond and beneath the historic church carrying both interior and exterior wood of Hemse. What more could we retrieve properties combined into a whole, we get to from this old archaeological material? simulate how these interact. This means that In this article we have pursued Emil Ekhoff we can experiment with effects of the sunlight and Olof Sörling through published texts and on the interior space and see how windows archival records. We may conclude that this and openings filter the outside world. one single building provides a large amount of data that they, at the time and with the means at their disposal, had difficulties seeing through, From reconstruction and furthermore, that their interpretation was embedded in a discourse of its time. to simulation Historical knowledge is, to quote Robin At the National History Museum’s exposition, George Collingwood, “the re-enactment of by the wall-mounted stave church facade with a past thought encapsulated in a context of original parts from Hemse, there is a sign that present thoughts” (Collingwood 1999). We informs the visitor: have revealed that the thoughts of Hemse intersected with the prestige and claim of the The earliest churches, such as the 11th- stave church brand. The appearance of Hemse century stave church in Hemse, were in the reconstruction drawing of Sörling and built of wood. Often they were built for Ekhoff is tainted by the discourse of its time. a magnate, on his own estate. During the The Norwegian connections are visualized 12th century an ecclesiastical organization by Urnes round corner staves and decorated took shape, with a church in every parish west gable and Hopperstad’s location of the and a cathedral in every diocese. Churches gable end decoration. In succeeding research

22 GUNNAR ALMEVIK & JONATHAN WESTIN Hemse was assigned to a transitional type of sides where the visitor has to choose one. stave church, as a link between the archaic The interactive model of Hemse palisade churches of Lund and the Norwegian presented in this research is not a definitive stave churches proper. reconstruction with inviolable properties. The From a forensic perspective we have re- interactive technology provides a means to examined the historic wood from Hemse present the alternatives and the ambiguities as and interrogated the accumulated sources for well. Reconstructing Hemse is the active part Gotlandic stave churches, and presented an of what Favro calls a knowledge representation alternative reconstruction model of the stave (Favro 2006): a process of knowledge church. The model shows the construction acquisition and evaluation that manifests and architecture of both exterior and itself in a reconstruction. Hence, the process interior space. We may point at the various of extracting information from a model being elements of the model, and argue for the developed constitutes knowledge acquisition visualized properties, one by one through by itself (Murteira et al. 2017). Moreover, primary sources or contextual analogies. The as Silberman writes, reconstruction “is not investigation also informs about events in the a conservation approach but an engagement life course of the building, that it has been approach that can help reconnect people with extensively repaired or moved and re-erected. place, history, and landscape” (Silberman The interactive model provides the 2015, 5). As such, the reconstruction is less of possibility for a dwelling perspective. The a static representation of our knowledge than forensic investigation and reconstruction a simulation or a research laboratory through may give the artefacts a presence effect, which which hypotheses can be tested and both connects cultural-historical interpretations researchers and the public can be engaged in to the effect of historic artefacts on our a dialogue. bodies and senses. The archaic appearance of a palisade church, with earth-bound wall staves fixed with stone boulders, although in References a construction with sill and top plate frames, Ahrens, C. 1981. Fruhe Holzkirchen im nördlichen seems like an “iteration” in conceptualizing Europa. Veröffentlichungen des Helms-Muse- a church building (Derrida 1971). The sill um 39. Hamburg. may have been a new constructive element Almevik, G. 2012. Byggnaden som kunskapskälla. but it was not integrated in the aesthetic Diss. Studies in Conservation, University of concept. Furthermore, the small church gives Gothenburg, Gothenburg. the impression of a kind of grandeur, with – 2017. Det forensiska perspektivet, In Almevik, G. (ed.), Hantverksvetenskap. Craft Laboratory, the inclination of the facades, the steep roof University of Gothenburg, Gothenburg. and the tall western portal. The main light Almevik, G. & Melin, K.-M. 2016. Conservation in the small church space, with a nave less Theory for Enhanced Craft Practice. In Van than 40 square metres, came from the main Balen, K. & Verstrynge, E. (eds.), Structural door. The centre column that is proposed Analysis of Historical Constructions. Anamnesis, Diagnosis, Therapy, Controls. Leuven. is not interpreted as a secondary support, Åman, A. & Järnfeldt-Carlsson, M. 1999. Träkyr- rather a primary constructive element and kor i Sverige. Stockholm. a distinguished social actant. The column Andrén, A. 2009. Vem lät bygga kyrkorna på Got- creates a network with the font, the side land? Saga och sed, Kungl. Gustav Adolfs akade- benches and the shrines. The column creates a miens årsbok. Annales Academiae Regiae Gus- frontline, a place before and beyond, and two tavi Adolphi. .

LUND ARCHAEOLOGICAL REVIEW 23 | 2017 23 Anker, L. 2016. Mellom tempel og korthus. In & Campana, S. (eds.), 3D Recording and Mo- Bakken, K. (ed.), Bevaring av stavkirkene, . delling in Archaeology and Cultural Heritage. – 1999. Gotlandske Stavkirker, Innlegg Jan Svan- Theory and Best Practices. Oxford. bergs Gotlandsseminar våren 1999. GF:s arkiv. Gibson, J. J. 2015[1979]. The Ecological Approach (Opublicerat småtryck) to Visual Perception. New York. Anker, P. 1997. Stavkirkene. Deres egenart og his- Ginzburg, C. & Davin, A. 1980. Morelli, Freud torie. Oslo. and Sherlock Holmes. Clues and Scientific Bartholin, T. S. 1998a. Dendrokronologi på Got- Method. History Workshop 9. land. Duplic. : . Gumbrecht, H. U. 2004. Production of Presence. Bartholin, T. S. 1998b. Dendrokronologiens til- What Meaning Cannot Convey. Stanford. forlidelighed. Fornvännen 93:2. Hansen, E. 2000. Fra malerisk til analytisk – op- Baudrillard, J. 1994. Simulacra and simulation. målingens historie i Danmark. In Sjömar P. Ann Arbor. (ed.), Byggnadsuppmätning, historik och prak- Bengtsson Melin, P. (2014) Medeltiden on Display. tik. Riksantikvarieämbetet, Stockholm. Uppställningar och utställningar av äldre kyrko- Haugli, R. 1976. Norske Stavkyrkor vols. I–II. konst från omkring 1850 till idag. Stockholm: Oslo. Runica et Mediaevalia, Centrum för medel- Ingold, T. 2011, The Perception of the Environment. tidsstudier Essays on Livelihood, Dwelling and Skill. Lon- Bjerknes, K. 1972. Stavkirke. In Kulturhistoriskt don. lexikon för nordisk medeltid bd. 17. Malmö. – 2007. Writing Texts, Reading Materials. A Re- Boëthius, G. 1931. Hallar, tempel och stavkyrkor. sponse to My Critics. Archaeological Dialogues Stockholm. 14(1). Bråthen, A. 1995. Dated wood from Gotland and Jørgensen, H., Lohfert, H., Laugerud, H. & Skin- the diocese of Skara. Højbjerg. nebach, L. K. (eds.), 2015. The Saturated Sen- Collingwood, R. G. 1999. The principles of history, sorium. Principles of Perception and Mediation and other writings in philosophy of history. Ox- in the . . ford. Lagerlöf, E. 1978. Gotlandskyrkornas takbelaggning. Day, J. (ed.) 2013. Making Senses of the Past. Vard och underhall av gotlandskyrkor. Toward a Sensory Archaeology. Carbondale. Riksantikvarieambetet och statens historiska De Reu, J., Plets, G., Verhoeven, G., De Smedt, museer, rapport 1978:3. Stockholm. P., Bats, M., Cherretté, B., De Maeyer, W., – 1972. Garde kyrka. Sveriges Kyrkor, vol. V:3. Deconynck, J., Herremans, D., Laloo, P., Van Stockholm. Meirvenne, M., De Clercq, W. 2013. Towards Lagerlöf, E. & Stolt, B. 1969. Hemse kyrkor. Sveri- a three-dimensional cost-effective registration ges Kyrkor vol. VI:3. Stockholm. of the archaeological heritage. Journal of Archa- – 1974. Eke kyrka. Sveriges Kyrkor vol. VI:6. eological Science 40(2). Stockholm. Derrida, J. 1988. Limited Inc (including “Signa- Lagerlöf, E., & Svahnstrom, G. 1991. Gotlands ture Event Context”). Evanston. kyrkor. Stockholm. Dmitruk, M. 2017. The town that never existed. Mogren, M., Roslund, M., Sundnér, B. & Wien- In Westin, J. & Martins Holmberg, I. (eds.), berg, J. (eds.) 2009. Triangulering. Historisk Memories of a City. Gothenburg. arkeologi vidgar fälten. Studies in Historical Ekhoff, E. 1914–16. Svenska stavkyrkor. Stockholm Archaeology 11. Lund. – 1896. En nyfunnen stafkyrka vid Hemse på Mol, A. 2002. The Body Multiple. Ontology in Gotland. Ny illustrerad tidning. Medical Practice. Durham. Fortidsminneforeningen 1899. Om den rette brug Murteira, H., Gago da Câmara, A., Simões Rod- av betegnelsen stavkirke. Published correspon- rigues, P. & Seueira, L., 2017. Lost Cities as dence between Herman Shirmer, Emil Ekhoff Virtual Experience. The Example of Pre-Earth- and Nicolay Nicolaysen) Fortidsminneforening- quake Lisbon. In Westin, J. and Martins Holm- ens årbok 1899, Oslo. berg, I. (eds.), Memories of a City. Gothenburg. Foucault, M. 2002. Vetandets arkeologi. [new ed.] Norberg-Schulz, C. 1980. Genius Loci: Towards a Lund. Phenomenology of Architecture. New York. Forte, M. 2014. Virtual Reality, Cyberarchaeology, Roosval, J. 1924. Den Baltiska Nordens kyrkor. Teleimmersive Archaeology. In Remondino, F. Uppsala.

24 GUNNAR ALMEVIK & JONATHAN WESTIN Silberman, N. 2015. Light at the End of the Laby- Archives rinth? From Historic Preservation to Heritage Forskningsföretaget Sveriges kyrkor Placemaking. New Approaches to the Inter- Primärmaterial för publicerade monografier Got- pretation of Historical Authenticity, Urban land F1ae, Vol. 8. 1961–1970 Conservation and Reconstruction in the Arabian Hemse stavkyrka, Hemse (vol. 131, Bd VI:3) Gulf. Amherst: University of Massachusetts. Smedberg, G. 1973. Nordens första kyrkor. En kyr- korättslig studie. Lund. Nationellt laboratorium för vedanatomi och Stolt, B. 2013. Den medeltida gudstjänstens till- dendrokronologi behör. Gotländskt arkiv 85. Dendrokronologiska undersökningar Gotland, Trotzig, G. 1972. En stavkyrka i Silte, Gotländskt Hemse stavkyrka (report 1998.02.03, T. S. arkiv 44. Bartholin, sample 12495-12505) Utas, J. 2008. Medeltida spåntak – fynd från got- ländska kyrkor. Byggnadshyttans årsskrift 2007– Emil Eckhoffs arkiv 2008. Visby: Byggnadshyttan på Gotland. Vol. 42, Svenska stavkyrkor, Anteckningar och Westin, J. 2014. Inking a Past. Visualization as a excerpter, 1. Shedding of Uncertainty. Visual Anthropology Vol. 43, Svenska stavkyrkor, Anteckningar och Review 30(2). excerpter, 2. Westin, J. & Almevik, G. 2017. A Virtual Dio- Vol. 44, Svenska stavkyrkor, Anteckningar och rama. Journal Nordic Museology. excerpter, 3. Vol. 45, Svenska stavkyrkor, Anteckningar och excerpter, 4. Vol. 62, 1900, Uppmätningsritningar

Gunnar Almevik, fil.dr, The Royal Swedish Academy of Letters, History and Antiquities research fellow, [email protected] Jonathan Westin, fil.dr, research fellow,[email protected] University of Gothenburg, Department of Conservation, Box 130, 405 30 Gothenburg, Sweden

LUND ARCHAEOLOGICAL REVIEW 23 | 2017 25 26